Error Checking..Please help! (1 Viewer)

[barnsey88]

Could someone please go over the error checking methods as I am really battling to understand them. The text book is so confusing. Thanks

 

[gman03]

Could someone please go over the error checking methods as I am really battling to understand them. The text book is so confusing. Thanks

I don't have the syllabus with me so i am guess which methods they are talking about


Parity:
Essentially counts the number of '1' in the binary form of the data.

Odd parity - odd number of '1' in data means correct transmission.
Similar for even parity


Checksum (need to check if it is correct):
Mathematically we add the individual bytes together, divide by a number and look at the remainder. If the remainder is, say 0, then the transmission is assumed proper.


CRC:
We start with a number x, then from the data every time we see a '0' we shift the number x to left. (i.e. multiply by two). Every time we see a '1' we xor the number with another number.

The point of CRC is that if there are random error the number x changes significantly.

 

[DigitalFortress]

CRC:
We start with a number x, then from the data every time we see a '0' we shift the number x to left. (i.e. multiply by two). Every time we see a '1' we xor the number with another number.

The point of CRC is that if there are random error the number x changes significantly.

what the fuck!
I thought CRC was division of the block by a 16bit or 32bit integer to get a massif floating point number.
This number is sent along with the block of data (similar to how checksum does it) then the recieving does the same calculation and bam, check final block of data with the recieving ends result and any discrepencies (its like a floating point number with massif precision) means an error has occured.

gman where'd u get that definition?

 

[gman03]

what the fuck!
I thought CRC was division of the block by a 16bit or 32bit integer to get a massif floating point number.
This number is sent along with the block of data (similar to how checksum does it) then the recieving does the same calculation and bam, check final block of data with the recieving ends result and any discrepencies (its like a floating point number with massif precision) means an error has occured.

gman where'd u get that definition?

http://en.wikipedia.org/wiki/Cyclic_redundancy_check

CRC can be done using pure integers. THat how I did it in Visual Basic 2 years ago.

Actually floating point is not needed in CRC32. It is true that CRC is a big massive division and what I described is one way of computing it.

Technically CRC is a high degree polynomial division.

 

[DigitalFortress]

shame hsc markers don't give a crap about real life :S

Text books make it a bit more simpler.